[1] We conducted a field experiment in a cool-temperate deciduous forest to investigate the dynamic behavior of soil CO2 and the vertical distribution of soil respiration. Soil CO2 concentration (C) was measured half-hourly at four depths for 6 months in 2000 with infrared gas analyzers installed below ground. Using C profiles, soil surface CO2 efflux (F-0), CO2 production rates of the topsoil (P-A), and CO2 flux from the subsoil to topsoil (F-CA) were evaluated half-hourly by applying Fick's first law. Some remarkable short-term and long-term variations were found in C, F-0, P-A, F-CA, and the contribution of topsoil respiration to total soil respiration (P-A/F-0), which include (1) rapid increase in C and decrease in F-0 and P-A due to rainwater infiltration, (2) diurnal variation in C coupled with that of the atmosphere, (3) diurnal variation in F-0 and P-A similar to that of topsoil temperature, (4) decrease in C, F-0, and P-A following soil drying in August, (5) linearly increasing F-CA between late May and mid-September, and (6) decrease in P-A/F-0 from around 0.9 during summer to 0.3 in November. The variation of P-A was mainly controlled by soil temperature at -0.07 m between 7degrees and 17degreesC, although P-A did not respond well to soil temperature above and below this temperature range. Above 17degreesC, P-A increased linearly with soil moisture, and moisture variation accounted for the P-A decrease in August. Neither temperature nor moisture explained the P-A behavior below 7degreesC. Subsoil respiration (F-CA) showed an exponential relationship with soil temperature at -1 m.